Phase reconstruction based on recurrent phase unwrapping with deep neural networks

Yoshiki Masuyama, Kohei Yatabe, Yuma Koizumi, Yasuhiro Oikawa, Noboru Harada

Research output: Contribution to journalArticlepeer-review

Abstract

Phase reconstruction, which estimates phase from a given amplitude spectrogram, is an active research field in acoustical signal processing with many applications including audio synthesis. To take advantage of rich knowledge from data, several studies presented deep neural network (DNN)-based phase reconstruction methods. However, the training of a DNN for phase reconstruction is not an easy task because phase is sensitive to the shift of a waveform. To overcome this problem, we propose a DNN-based two-stage phase reconstruction method. In the proposed method, DNNs estimate phase derivatives instead of phase itself, which allows us to avoid the sensitivity problem. Then, phase is recursively estimated based on the estimated derivatives, which is named recurrent phase unwrapping (RPU). The experimental results confirm that the proposed method outperformed the direct phase estimation by a DNN.

Original languageEnglish
JournalUnknown Journal
Publication statusPublished - 2020 Feb 13

Keywords

  • Group delay
  • Instantaneous frequency
  • Recurrent neural network
  • Spectrogram inversion
  • Time-frequency analysis

ASJC Scopus subject areas

  • General

Fingerprint Dive into the research topics of 'Phase reconstruction based on recurrent phase unwrapping with deep neural networks'. Together they form a unique fingerprint.

Cite this